kyuubi/docs/deployment/settings.md

Introduction to the Kyuubi Configurations System

Kyuubi provides several ways to configure the system and corresponding engines.

Environments

You can configure the environment variables in $KYUUBI_HOME/conf/kyuubi-env.sh, e.g, JAVA_HOME, then this java runtime will be used both for Kyuubi server instance and the applications it launches. You can also change the variable in the subprocess's env configuration file, e.g.$SPARK_HOME/conf/spark-env.sh to use more specific ENV for SQL engine applications.

#!/usr/bin/env bash
#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
#
# - JAVA_HOME               Java runtime to use. By default use "java" from PATH.
#
#
# - KYUUBI_CONF_DIR         Directory containing the Kyuubi configurations to use.
#                           (Default: $KYUUBI_HOME/conf)
# - KYUUBI_LOG_DIR          Directory for Kyuubi server-side logs.
#                           (Default: $KYUUBI_HOME/logs)
# - KYUUBI_PID_DIR          Directory stores the Kyuubi instance pid file.
#                           (Default: $KYUUBI_HOME/pid)
# - KYUUBI_MAX_LOG_FILES    Maximum number of Kyuubi server logs can rotate to.
#                           (Default: 5)
# - KYUUBI_JAVA_OPTS        JVM options for the Kyuubi server itself in the form "-Dx=y".
#                           (Default: none).
# - KYUUBI_NICENESS         The scheduling priority for Kyuubi server.
#                           (Default: 0)
# - KYUUBI_WORK_DIR_ROOT    Root directory for launching sql engine applications.
#                           (Default: $KYUUBI_HOME/work)
# - HADOOP_CONF_DIR         Directory containing the Hadoop / YARN configuration to use.
#
# - SPARK_HOME              Spark distribution which you would like to use in Kyuubi.
# - SPARK_CONF_DIR          Optional directory where the Spark configuration lives.
#                           (Default: $SPARK_HOME/conf)
#


## Examples ##

# export JAVA_HOME=/usr/jdk64/jdk1.8.0_152
# export HADOOP_CONF_DIR=/usr/ndp/current/mapreduce_client/conf
# export KYUUBI_JAVA_OPTS="-Xmx10g -XX:+UnlockDiagnosticVMOptions -XX:ParGCCardsPerStrideChunk=4096 -XX:+UseParNewGC -XX:+UseConcMarkSweepGC -XX:+CMSConcurrentMTEnabled -XX:CMSInitiatingOccupancyFraction=70 -XX:+UseCMSInitiatingOccupancyOnly -XX:+CMSClassUnloadingEnabled -XX:+CMSParallelRemarkEnabled -XX:+UseCondCardMark -XX:MaxDirectMemorySize=1024m  -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=./logs -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCDateStamps -XX:+PrintTenuringDistribution -Xloggc:./logs/kyuubi-server-gc-%t.log -XX:+UseGCLogFileRotation -XX:NumberOfGCLogFiles=10 -XX:GCLogFileSize=5M -XX:NewRatio=3 -XX:MetaspaceSize=512m"

Kyuubi Configurations

You can configure the Kyuubi properties in $KYUUBI_HOME/conf/kyuubi-defaults.conf. For example:

#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#

## Kyuubi Configurations

#
# kyuubi.authentication           NONE
# kyuubi.frontend.bind.host       localhost
# kyuubi.frontend.bind.port       10009
#

# Details in https://kyuubi.readthedocs.io/en/latest/deployment/settings.html

Authentication

Key	Default	Meaning	Since
kyuubi.authentication	NONE	Client authentication types. NOSASL: raw transport. NONE: no authentication check. KERBEROS: Kerberos/GSSAPI authentication. LDAP: Lightweight Directory Access Protocol authentication.	1.0.0
kyuubi.authentication .ldap.base.dn	<undefined>	LDAP base DN.	1.0.0
kyuubi.authentication .ldap.domain	<undefined>	LDAP domain.	1.0.0
kyuubi.authentication .ldap.url	<undefined>	SPACE character separated LDAP connection URL(s).	1.0.0
kyuubi.authentication .sasl.qop	auth	Sasl QOP enable higher levels of protection for Kyuubi communication with clients. auth - authentication only (default) auth-int - authentication plus integrity protection auth-conf - authentication plus integrity and confidentiality protection. This is applicable only if Kyuubi is configured to use Kerberos authentication.	1.0.0

Backend

Key	Default	Meaning	Since
kyuubi.backend.engine .exec.pool.keepalive .time	PT1M	Time(ms) that an idle async thread of the operation execution thread pool will wait for a new task to arrive before terminating in SQL engine applications	1.0.0
kyuubi.backend.engine .exec.pool.shutdown .timeout	PT10S	Timeout(ms) for the operation execution thread pool to terminate in SQL engine applications	1.0.0
kyuubi.backend.engine .exec.pool.size	100	Number of threads in the operation execution thread pool of SQL engine applications	1.0.0
kyuubi.backend.engine .exec.pool.wait.queue .size	100	Size of the wait queue for the operation execution thread pool in SQL engine applications	1.0.0
kyuubi.backend.server .exec.pool.keepalive .time	PT1M	Time(ms) that an idle async thread of the operation execution thread pool will wait for a new task to arrive before terminating in Kyuubi server	1.0.0
kyuubi.backend.server .exec.pool.shutdown .timeout	PT10S	Timeout(ms) for the operation execution thread pool to terminate in Kyuubi server	1.0.0
kyuubi.backend.server .exec.pool.size	100	Number of threads in the operation execution thread pool of Kyuubi server	1.0.0
kyuubi.backend.server .exec.pool.wait.queue .size	100	Size of the wait queue for the operation execution thread pool of Kyuubi server	1.0.0

Delegation

Key	Default	Meaning	Since
kyuubi.delegation.key .update.interval	PT24H	unused yet	1.0.0
kyuubi.delegation .token.gc.interval	PT1H	unused yet	1.0.0
kyuubi.delegation .token.max.lifetime	PT168H	unused yet	1.0.0
kyuubi.delegation .token.renew.interval	PT168H	unused yet	1.0.0

Frontend

Key	Default	Meaning	Since
kyuubi.frontend .backoff.slot.length	PT0.1S	Time to back off during login to the frontend service.	1.0.0
kyuubi.frontend.bind .host	<undefined>	Hostname or IP of the machine on which to run the frontend service.	1.0.0
kyuubi.frontend.bind .port	10009	Port of the machine on which to run the frontend service.	1.0.0
kyuubi.frontend.login .timeout	PT20S	Timeout for Thrift clients during login to the frontend service.	1.0.0
kyuubi.frontend.max .message.size	104857600	Maximum message size in bytes a Kyuubi server will accept.	1.0.0
kyuubi.frontend.max .worker.threads	999	Maximum number of threads in the of frontend worker thread pool for the frontend service	1.0.0
kyuubi.frontend.min .worker.threads	9	Minimum number of threads in the of frontend worker thread pool for the frontend service	1.0.0
kyuubi.frontend .worker.keepalive.time	PT1M	Keep-alive time (in milliseconds) for an idle worker thread	1.0.0

Ha

Key	Default	Meaning	Since
kyuubi.ha.zookeeper .acl.enabled	false	Set to true if the zookeeper ensemble is kerberized	1.0.0
kyuubi.ha.zookeeper .connection.base.retry .wait	1000	Initial amount of time to wait between retries to the zookeeper ensemble	1.0.0
kyuubi.ha.zookeeper .connection.max .retries	3	Max retry times for connecting to the zookeeper ensemble	1.0.0
kyuubi.ha.zookeeper .connection.max.retry .wait	30000	Max amount of time to wait between retries for BOUNDED_EXPONENTIAL_BACKOFF policy can reach, or max time until elapsed for UNTIL_ELAPSED policy to connect the zookeeper ensemble	1.0.0
kyuubi.ha.zookeeper .connection.retry .policy	EXPONENTIAL_BACKOFF	The retry policy for connecting to the zookeeper ensemble, all candidates are: ONE_TIME N_TIME EXPONENTIAL_BACKOFF BOUNDED_EXPONENTIAL_BACKOFF UNTIL_ELAPSED	1.0.0
kyuubi.ha.zookeeper .connection.timeout	15000	The timeout(ms) of creating the connection to the zookeeper ensemble	1.0.0
kyuubi.ha.zookeeper .namespace	kyuubi	The root directory for the service to deploy its instance uri. Additionally, it will creates a -[username] suffixed root directory for each application	1.0.0
kyuubi.ha.zookeeper .quorum		The connection string for the zookeeper ensemble	1.0.0
kyuubi.ha.zookeeper .session.timeout	60000	The timeout(ms) of a connected session to be idled	1.0.0

Kinit

Key	Default	Meaning	Since
kyuubi.kinit.interval	PT1H	How often will Kyuubi server run kinit -kt [keytab] [principal] to renew the local Kerberos credentials cache	1.0.0
kyuubi.kinit.keytab	<undefined>	Location of Kyuubi server's keytab.	1.0.0
kyuubi.kinit.max .attempts	10	How many times will kinit process retry	1.0.0
kyuubi.kinit .principal	<undefined>	Name of the Kerberos principal.	1.0.0

Metrics

Key	Default	Meaning	Since
kyuubi.metrics .enabled	true	Set to true to enable kyuubi metrics system	1.2.0
kyuubi.metrics.json .report.location	metrics	Where the json metrics file located	1.2.0
kyuubi.metrics.report .interval	PT5S	How often should report metrics to json/console. no effect on JMX	1.2.0
kyuubi.metrics .reporters	JSON	A comma separated list for all metrics reporters JSON - default reporter which outputs measurements to json file periodically CONSOLE - ConsoleReporter which outputs measurements to CONSOLE. SLF4J - Slf4jReporter which outputs measurements to system log. JMX - JmxReporter which listens for new metrics and exposes them as namespaced MBeans.	1.2.0

Operation

Key	Default	Meaning	Since
kyuubi.operation.idle .timeout	PT3H	Operation will be closed when it's not accessed for this duration of time	1.0.0
kyuubi.operation .interrupt.on.cancel	true	When true, all running tasks will be interrupted if one cancels a query. When false, all running tasks will remain until finished.	1.2.0
kyuubi.operation .query.timeout	PT0S	Set a query duration timeout in seconds in Kyuubi. If the timeout is set to a positive value, a running query will be cancelled automatically if timeout. Otherwise the query continues to run till completion. If timeout values are set for each statement via java.sql.Statement.setQueryTimeout and they are smaller than this configuration value, they take precedence. If you set this timeout and prefer to cancel the queries right away without waiting task to finish, consider enabling kyuubi.operation.interrupt.on.cancel together.	1.2.0
kyuubi.operation .status.polling .timeout	PT5S	Timeout(ms) for long polling asynchronous running sql query's status	1.0.0

Session

Key	Default	Meaning	Since
kyuubi.session.check .interval	PT5M	The check interval for session timeout.	1.0.0
kyuubi.session.engine .check.interval	PT5M	The check interval for engine timeout	1.0.0
kyuubi.session.engine .idle.timeout	PT30M	engine timeout, the engine will self-terminate when it's not accessed for this duration	1.0.0
kyuubi.session.engine .initialize.timeout	PT1M	Timeout for starting the background engine, e.g. SparkSQLEngine.	1.0.0
kyuubi.session.engine .log.timeout	PT24H	If we use Spark as the engine then the session submit log is the console output of spark-submit. We will retain the session submit log until over the config value.	1.1.0
kyuubi.session.engine .login.timeout	PT15S	The timeout(ms) of creating the connection to remote sql query engine	1.0.0
kyuubi.session.engine .share.level	USER	The SQL engine App will be shared in different levels, available configs are: CONNECTION: the App will not be shared but only used by the current client connection USER: the App will be shared by all sessions created by a unique username SERVER: the App will be shared by Kyuubi servers	1.0.0
kyuubi.session.engine .spark.main.resource	<undefined>	The package used to create Spark SQL engine remote application. If it is undefined, Kyuubi will use the default	1.0.0
kyuubi.session.engine .startup.error.max .size	8192	During engine bootstrapping, if error occurs, using this config to limit the length error message(characters).	1.1.0
kyuubi.session .timeout	PT6H	session timeout, it will be closed when it's not accessed for this duration	1.0.0

Zookeeper

Key	Default	Meaning	Since
kyuubi.zookeeper .embedded.directory	embedded_zookeeper	The temporary directory for the embedded zookeeper server	1.0.0
kyuubi.zookeeper .embedded.port	2181	The port of the embedded zookeeper server	1.0.0

Spark Configurations

Via spark-defaults.conf

Setting them in $SPARK_HOME/conf/spark-defaults.conf supplies with default values for SQL engine application. Available properties can be found at Spark official online documentation for Spark Configurations

Via kyuubi-defaults.conf

Setting them in $KYUUBI_HOME/conf/kyuubi-defaults.conf supplies with default values for SQL engine application too. These properties will override all settings in $SPARK_HOME/conf/spark-defaults.conf

Via JDBC Connection URL

Setting them in the JDBC Connection URL supplies session-specific for each SQL engine. For example: jdbc:hive2://localhost:10009/default;#spark.sql.shuffle.partitions=2;spark.executor.memory=5g

• Runtime SQL Configuration
  • For Runtime SQL Configurations, they will take affect every time
• Static SQL and Spark Core Configuration
  • For Static SQL Configurations and other spark core configs, e.g. spark.executor.memory, they will take affect if there is no existing SQL engine application. Otherwise, they will just be ignored

Via SET Syntax

Please refer to the Spark official online documentation for SET Command

Logging

Kyuubi uses log4j for logging. You can configure it using $KYUUBI_HOME/conf/log4j.properties.

#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#

# Set everything to be logged to the console
log4j.rootCategory=INFO, console
log4j.appender.console=org.apache.log4j.ConsoleAppender
log4j.appender.console.target=System.err
log4j.appender.console.layout=org.apache.log4j.PatternLayout
log4j.appender.console.layout.ConversionPattern=%d{yyyy-MM-dd HH:mm:ss.SSS} %p %c{2}: %m%n

Other Configurations

Hadoop Configurations

Specifying HADOOP_CONF_DIR to the directory contains hadoop configuration files or treating them as Spark properties with a spark.hadoop. prefix. Please refer to the Spark official online documentation for Inheriting Hadoop Cluster Configuration. Also, please refer to the Apache Hadoop's online documentation for an overview on how to configure Hadoop.

Hive Configurations

These configurations are used for SQL engine application to talk to Hive MetaStore and could be configured in a hive-site.xml. Placed it in $SPARK_HOME/conf directory, or treating them as Spark properties with a spark.hadoop. prefix.

User Defaults

In Kyuubi, we can configure user default settings to meet separate needs. These user defaults override system defaults, but will be overridden by those from JDBC Connection URL or Set Command if could be. They will take effect when creating the SQL engine application ONLY.

User default settings are in the form of ___{username}___.{config key}. There are three continuous underscores(_) at both sides of the username and a dot(.) that separates the config key and the prefix. For example:

# For system defaults
spark.master=local
spark.sql.adaptive.enabled=true
# For a user named kent
___kent___.spark.master=yarn
___kent___.spark.sql.adaptive.enabled=false
# For a user named bob
___bob___.spark.master=spark://master:7077
___bob___.spark.executor.memory=8g

In the above case, if there are related configurations from JDBC Connection URL, kent will run his SQL engine application on YARN and prefer the Spark AQE to be off, while bob will activate his SQL engine application on a Spark standalone cluster with 8g heap memory for each executor and obey the Spark AQE behavior of Kyuubi system default. On the other hand, for those users who do not have custom configurations will use system defaults.